Pancreatic -cells expressing GLP-1 are resistant to the toxic effects of immunosuppressive drugs

Pancreatic �-cells expressing GLP-1 are resistant to the toxiceffects of immunosuppressive drugs

Eugenio D’Amico1,2, Hongxiang Hui1,3, Nasif Khoury1, Umberto Di Mario4 andRiccardo Perfetti1,3

1Division of Endocrinology, Cedars-Sinai Medical Center, Los Angeles, California, USA

2Universita’ ‘Magna Graecia’, Catanzaro, Italy

3University of California, Los Angeles, California, USA

4Department of Clinical Sciences, University ‘La Sapienza’, Rome, Italy

(Requests for offprints should be addressed to R Perfetti; Email: [email protected])

Abstract

Glucose intolerance is often observed after pancreatic islet cell transplantation. The administration ofimmunosuppressive agents (ISD), necessary to avoid tissue rejection, is in part responsible for hyperglycemia. Toinvestigate whether mouse insulinoma (MIN6) cells transfected with the glucagon like peptide-1 (GLP-1) fragment ofthe proglucagon gene (RIP/GLP-1 MIN6 cells) are resistant to the toxicity derived from the administration of ISD.RIP/GLP-1 MIN6 cells, as well as parental MIN6 cells, were exposed to a cocktail of ISD. The secretion of insulin andthe expression of apoptosis-related proteins were investigated by RIA and western blot analysis. Cell apoptosis wasquantified by FACS analysis. Finally, to study whether the antiapoptotic action of GLP-1 was a function of its effect oninsulin secretion, or rather it was a direct effect of GLP-1, cells were cultured with or without diazoxide or exendin-9.GLP-1 improved the functional activity and the viability of cells exposed to ISD. The insulin secretion of RIP/GLP-1MIN6 cells after exposure to ISD was preserved. The expression of GLP-1 by �-cells reduced the number of apoptoticcells and increased the expression of the antiapoptotic protein Bcl-2. GLP-1 also decreased the abundance of theproapoptotic markers PARP-p85 and Smac/Diablo. Treatment of cells with the diazoxide did not abolish the protectiveadvantage that cells transfected with GLP-1 had; conversely the exposure of cells to exendin-9 was associated with arestored susceptibility to apoptosis. This report demonstrates that GLP-1 is capable of preserving �-cell function andprotecting cells from apoptotic cell death.

Journal of Molecular Endocrinology (2005) 34, 377–390

Introduction

The control of plasma glucose levels has beenrecognized to be a major determinant in preventingand/or delaying the onset of complications in patientswith diabetes (Lakey et al. 1999). Equally well establishedis the observation that maintaining normal blood glucoseconcentrations is often quite a difficult goal to reach.The transplantation of pancreatic tissue, either thewhole organ or purified pancreatic islets, has led to arenewed interest in transplantation as a treatment fordiabetes mellitus (The Diabetes Control and Com-plications Trial Research Group 1998, Drachenberget al. 1999). With the increase in the availability of newand more potent immunosuppressive drugs, differentstrategies have been proposed specifically for islettransplantation. Shapiro et al. (2000) have developed aglucocorticoid-free immunosuppressive protocol thatincludes sirolimus, low-dose tacrolimus, and a mono-clonal antibody against the interleukin-2 receptor(daclizumab) for islet transplantation in patients with

brittle type 1 diabetes. The addition of mycophenolatemofetil to immunosuppressive regimens has been shownto reduce even further the incidence of rejectionfollowing pancreas or islets transplantation (Sato et al.2003).

A recent review of the potential barriers to insulinindependence after islet transplantation identified severalfactors (Hering & Ricordi 1999). The number of �-cellsmay be inadequate owing to insufficient engraftment ofislets and immediate cellular loss through apoptosis andother non-immune-mediated inflammatory pathways(Bennet et al. 1999). A persistent impairment of glucosemetabolism after transplantation has been attributed tovarious causes, including impaired insulin secretion(Patty et al. 2002) and �-cells apoptosis induced by theadministration of immunosuppressive drugs (ISD)(Drachenberg et al. 1999).

For any type of immunossuppression after transplan-tation, a balance is sought between efficacy andtoxicity. A major advance toward achieving insulinindependence following islet transplantation or pancreas

377

Journal of Molecular Endocrinology (2005) 34, 377–3900952–5041/05/034–377 © 2005 Society for Endocrinology Printed in Great Britain

DOI: 10.1677/jme.1.01655Online version via http://www.endocrinology-journals.org

http://www.endocrinology-journals.org

transplantation is to eliminate, or minimize, the use ofimmunosuppressive agents known to have �-cell toxicity(Zeng et al. 1993) or to induce peripheral insulinresistance (Sutherland et al. 2001). An alternativestrategy might consider ways to enhance the resistanceof islets to the toxic effects of ISD.

Recent reports have shown that the gastrointestinalhormone glucagon-like peptide 1 (GLP-1), in addition toregulating the secretion of insulin and the expression ofislet-specific genes (Hussain & Habener 2000, Drucker2003), has significant antiapoptotic properties. Indiabetic ZDF rats, as well as in streptozotocin-induceddiabetic mice, treatment with GLP-1 or exendin-4 (along-acting analog of GLP-1) protects �-cells fromapoptosis (Farilla et al. 2002, Li et al. 2003). In vitro studiesusing insulin-secreting cells (Hui et al. 2002) orfibroblast-expressing GLP-1 receptors (Li et al. 2003)have shown that the protective action of GLP-1 is not anindirect effect resulting from the normalization of theambient glucose levels, but rather it has a direct effect oncell viability and is mediated by an increase inexpressionof antiapoptotic proteins and down-regulation ofproapoptotic proteins (Farilla et al. 2002, Hui et al. 2003,Li et al. 2003).

The aim of this study was to investigate whetherGLP-1 would confer resistance to the toxic effects ofISD, preserving the viability and function of �-cells.Using our previously described �-cell line transfectedwith the GLP-1 fragment of the human proglucagongene (Hui et al. 2002), we demonstrate that GLP-1counteracts the negative effects of ISD, improving theglucose-dependent capability to secrete insulin andreducing �-cell death.

Materials and methods

Cell culture

Mouse insulinoma (MIN-6) cells were a gift from DrJunichi Miyazaki (Kumamoto University; Kumamoto,Japan). Transfection of MIN6 cells with the GLP-1fragment of the human proglucagon gene, under thecontrol of the rat insulin II promoter (RIP/GLP-1construct) has been previously described by Hui et al.(2002). RIP/GLP-1 MIN6 cells have been shown to becapable of promoting the synthesis and secretion ofGLP-1 and insulin in a glucose-dependent manner. Cellswere cultured in Dulbecco’s modified Eagle’s medium(DMEM) medium (Life Technologies Inc., Gaithers-burg, MD, USA) containing100 µg/ml penicillin,50 µg/ml streptomycin, and 10% fetal bovine serum(FBS) (Gemini Bio-Products Inc., Woodland, CA, USA)at 37 �C under a humidified condition of 95% air and5% CO2. All cell culture studies were conducted usingcells grown to 80% confluence and exposed for 48 h tomedium containing a cocktail of ISD, or an equivalent

volume of the medium used to dilute each of the drugsused. The ISD used included sirolimus (Rapamycin,Sigma Aldrich, St Louis, MO, USA) at concentrations of25 ng/ml; mycophenolate, (Mycophenolic acid, Calbio-chem, La Jolla, CA, USA) at concentrations of 17·5ug/ml; and tacrolimus (Prograf, Fujisawa, Deerfield, NJ,USA) at concentrations of 75 ng/ml. Drug dilutionswere made from fresh stock before each experiment.Sirolimus was solubilized in dimethyl sulfoxide, myco-phenolate in 70% ethanol, and tacrolimus in sterilewater. The selection of those agents was based on theprotocol for immunosuppression often employed afterpancreatic islets transplantation (Shapiro et al. 2000), andthe concentrations of ISD used in the present study wereselected on the basis of previous experiments performedin our laboratory (Hui et al. 2004).

In some studies, human recombinant GLP-1(American Peptide Company; Sunnyvale, CA, USA) wasadded to the culture medium of parental MIN6 cells.Those experiments were aimed at investigating whetherthe effect of ISD on RIP/GLP-1 MIN6 cells wasspecifically due to their capability of producing GLP-1,or whether it was the effect of a clonal selection resultingfrom the process of cell transfection. Cells wereevaluated for insulin secretion, viability, and theexpression of anti-apoptotic and pro-apoptotic factors.The specific culture conditions for each individualexperiment are detailed in the specific experimentalprotocol described hereafter. All experiments withtransfected MIN6 cells were performed using pools ofstably transfected clones.

Measurement of insulin secretion

Parental MIN6 and RIP/GLP-1 MIN-6 cells wereplated at a density of 1�106 cells/well in 6-wells plates.After they reached confluence, cells were subjected to a2 h wash-out and cultured in serum-free mediumcontaining 25 mM glucose and a cocktail of ISD(sirolimus (25 ng/ml), mycophenolate (17·5 ug/ml) andtacrolimus (75 ng/ml)) for 48 h. Control cultures wereincubated in the absence of ISD for the same length oftime. After a 48 h treatment the cells were subjected to a2 h wash-out period in serum-free/glucose-free medium,and exposed for 1 h to medium containing either 5·6 or16·7 mM glucose. The level of insulin in the culturemedium was measured by RIA (Linco Research Inc.,St Charles, MA, USA) and normalized for the total cellularprotein content detected in the pellet of each individualculture, as measured using the Bradford method (Bio-RadLaboratories Inc., Richmond, CA, USA).

LIVE/DEAD viability/cytotoxicity

Parental and RIP/GLP-1 MIN6 cells were grown inchamber slides and treated with a cocktail of ISD and

E D’AMICO and others · Pancreatic �-cells expressing GLP-1 resistance to ISD378

www.endocrinology-journals.orgJournal of Molecular Endocrinology (2005) 34, 377–390


analyszed with the LIVE/DEAD Viability/CytotoxicityKit (Probes, Eugene, OR, USA). After 48 h treatment, thecells were collected, washed twice with PBS, stained for30 min, and visualized under UV light with a BH-2microscope (Olympus Corp., Melville, NY, USA). This isa fluorescence-based method for determining viability ofadherent or non-adherent cells and for assaying cytotoxic-ity. The assay comprises of two probes: calcein AM andethidium homodimer-1. Calcein AM is a fluorogenicesterase substrate that is hydrolyzed to a green-fluorescentproduct (calcein); thus, the emission of green fluorescentlight is an indicator of cells that have esterase activityas well as an intact membrane to retain the esteraseproducts. Ethidium homodimer-1 is a high-affinity, red-fluorescent nucleic acid stain that is only able to passthrough the compromised membranes of dead cells.

Immunostaining for insulin and the active form ofcaspase-3

After a 48 h treatment with or without ISD, cells werefixed for 1 h at room temperature in 4% paraform-aldehyde in PBS (pH 7·4). They were then washed inPBS, permeabilized in 0·1% Triton X-100 in 0·1%sodium citrate, rinsed twice in PBS, and incubated with apolyclonal goat anti active form of caspase-3 antibody(Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA)diluted 1:250 in PBS, or with a polyclonal guinea piganti-insulin antibody (DAKO Corp., Carpinteria, CA,USA) diluted 1:400. The incubation was performed at4 �C overnight. After several washes in PBS, an antigoatfluorescein-conjugated antibody (Probes, Eugene, OR,USA) diluted 1:1500 for the detection of caspase-3, andan anti-guinea pig rhodamine-conjugated antibody(Probes), diluted 1:1500 for the detection of insulin, wereincubated for 1 h at room temperature. To identify thenuclei, the cells were also stained with the karyophilic dyeHoechst 33342 (10 µ g/ml) for 5 min at room tempera-ture. After a final wash in PBS, the cells were visualizedunder UV light with a BH-2 microscope (OlympusCorp., Melville, NY, USA) and using a fluoresceinstandard filter set (520�20 nm) for caspase-3 stainingand a rhodamine filter (>620 nm) for insulin.

FACS analysis for Annexin-V

Apoptotic cells were analyzed by flow cytometer (FACS;Becton Dickinson and Co., San Jose, CA, USA) usingthe Annexin-V-FLOUS Staining Kit (Roche DiagnosticCorporation, Indianapolis, IN, USA). Annexin-V is aCa2+-dependent phospholipid-binding protein with ahigh affinity for phosphotidylserine (PS); hence, thisprotein can be used as a probe for PS exposure on theouter leaflet of the cell membrane and can be used forthe detection of apoptotic cells. The simultaneousapplication of propidium iodide as a DNA stain, used for

dye exclusion tests, allows Annexin-V positively stainedcells to be distinguished from necrotic cells. Cells werecollected from the culture flasks and washed twice withPBS by centrifugation at 200 g for 5 min. The cell pelletwas then suspended in 100 µl of staining solution andincubated for 15 min at room temperature. Flowcytometric analysis was performed with a FACScancytometer (Becton Dickinson and Co., Franklin Lakes,NJ, USA), using the LYSIS II analyzer program.Exendin (9–39) was purchased from American PeptideCo. (Sunnyvale, CA, USA).

DNA fragmentation analysis

Floating and adherent cells from each culture conditionstudied were collected, centrifuged, pelleted at 400 g for5 min, and washed twice with PBS. The pellet was thenresuspended in 0·2 ml lysis buffer (100 mM NaCl,10 mM Tris (pH 8·0), 1 mM EDTA, 0·5% sodiumdodecyl sulfate, 0·20 mg/ml proteinase K, 200 µg/mlribonuclease A). The cell lysates were incubated at 37 �Cfor 2 h. The genomic DNA was extracted by twoseparations with phenol/chloroform, followed by oneextraction with chloroform alone. The DNA pellet wasthen washed in 70% ethanol and resuspended in 1 mMEDTA and 10 mM Tris–HCl (pH 8·0) at a finalconcentration of 20 µg/ml. The DNA fragmentationanalysis was performed using a 1·5% agarose gel in1 mM EDTA, 40 mM Tris acetate (pH 7·6) to visualizethe laddering of the samples.

Western blot analysis

Cell pellets were lysed at 4 �C in a buffer containing60 mM Tris–HCl (pH 6·8), 1% sodium dodecyl sulfate,10% glycerol, 0·05% NP-40 and 0·5%�-mercaptoethanol and protease inhibitor mixture(1:100 dilution). The lysis buffer for the proteinextraction to determine the cytoplasmic abundance ofSmac was prepared with the same reagents listed above,without NP-40. Lysates were cleared at 12 000 r.p.m. for15 min at 4 �C and stored at �80 �C until used. Theprotein concentration was determined using a Comassiedye assay (Bio Rad, Hercules, CA, USA), and BSA(Sigma Aldrich) was used as a standard. The cell lysates(25 µg per sample) were then separated by 8% or 12%SDS-PAGE under reducing conditions and electrotrans-ferred onto Hybond-C nitrocellulose membrane (BioRad) using standard procedures. The membranes wereincubated for 2–4 h at room temperature with TBST(20 mM Tris–HCl (pH 7·5), 150 mM NaCl, 0·2%Tween-20) detection reagent. Primary antibodieswere used at the following dilutions: anti Bcl-2 (1:500);anti poly-(ADP-ribose)-polymerase (PARP; 1:500); antisecond mitochondria-derived activator of caspase Smac/DIABLO (1:400); anti active form of caspase-3 (1:1000);

Pancreatic �-cells expressing GLP-1 resistance to ISD · E D’AMICO and others 379

www.endocrinology-journals.org Journal of Molecular Endocrinology (2005) 34, 377–390


and anti �-actin (1:1000). All antibodies were purchasedfrom Santa Cruz Biotechnology Inc.

Measurement of GLP-1 secretion

Parental and RIP/GLP-1 MIN6 cells utilized to performFACS analysis for Annexin-V were also used to evaluatethe concentration of GLP-1 in their culture medium.GLP-1 was measured by RIA (Linco Research Inc.,St Charles, MA, USA).

Statistical analysis

The data were expressed as mean�S.E. Comparisons ofmeans of multiple groups were made using one-wayANOVA followed by post hoc analysis using Fisher’s leastsignificant difference method.

Results

GLP-1 counteracted the negative effects of ISD onglucose-dependent insulin secretion

The exposure of parental MIN6 cells to a cocktail ofISD for 48 h produced a significant reduction of their

ability to secrete insulin after glucose challenge. Indeed,when parental MIN6 cells were exposed (for 1 h) tomedium containing 16·7 mM glucose (after a 2 h washout with glucose-free, serum-free medium) there was asignificantly decrease in the insulin concentration incultures exposed to ISD (1·02�0·21 ng insulin/mgpellet protein) compared with cultures exposed to vehiclealone (1·71�0·19 ng insulin/mg pellet protein; P<0·01)(Fig. 1). Conversely, RIP/GLP-1 MIN6 cells showed asignificant preservation of glucose-dependent secretionof insulin after treatment with ISD. Indeed, exposure ofRIP/GLP-1 MIN6 cells to ISD was associated with anaccumulation of insulin in the culture medium that wasequivalent to approximately 85% of the total insulinaccumulated in the medium of untreated RIP/GLP-1MIN6 cells (Fig. 1).

GLP-1 preserved the viability of cells exposed toISD

Qualitative analysis of the effect of GLP-1 on cell viability

The viability of parental and RIP/GLP-1 MIN6 cellsexposed to either a cocktail of ISD or vehicle for 48 hwas measured using calcein-AM (a membrane-permeable dye that in live cells is cleaved by esterases to

Figure 1 Effect of ISD on glucose-stimulated insulin secretion: constitutive expression of GLP-1counteracts the negative effect of ISD on glucose-dependent insulin secretion. Parental andRIP/GLP-1 MIN6 cells were cultured for 48 h in the presence or absence of ISD (sirolimus (25 ng/ml),mycophenolate (17·5 ug/ml) and tacrolimus (75 ng/ml)). After a 2 h washout with medium (in theabsence of FBS and glucose), the cells were challenged with either 5·6 or 16·7 mM glucose. Eachexperiment was repeated at least four times, and the data plotted on the graph represent the mean±S.E.

Insulin levels were normalized for total proteins from cell extracts. *P,0·01; NS=not significant.




yield cytoplasmic green fluorescence) and ethidiumhomodimer-1 (a red nucleic acid dye to which live cellsare impermeable, but that accumulates in the nuclei ofmembrane-compromised cells) staining. Panel A and Bof Figure 2 show that in cultures of both parental MIN6cells and RIP/GLP-1 MIN6 cells grown without ISDvirtually all cells were viable. Both cultures showed alarge number of dead cells after exposure to ISD.However, cultures of parental MIN6 cells (C) showed agreater number of dead cells, compared with cultures ofRIP/GLP-1 MIN6 cells (D).

Quantitative analysis of the effect of GLP-1 on cell viability

FACS analysis for Annexin-V of parental and RIP/GLP-1 MIN6 cells exposed to either ISD or vehicleindicated that the expression of GLP-1 by cells enhancedthe percentage of viable cells (Table 1). This experimentindicated that 18·83�5% of parental MIN6 cells vs.8·33�5% of the RIP/GLP-1 MIN6 cells wereAnnexin-V-positive when exposed to ISD (P<0·01). Thetransfection of cells with GLP-1 was also associated withsignificant protection from necrosis induced by ISD

(P<0·01) and an overall larger percentage of living cells(75·64�5% vs. 58·51�5% in RIP/GLP-1 MIN6 cellsvs. parental MIN6, P<0·01), after challenge with ISD(P<0·01; Table 1).

RIA for GLP-1 of the culture medium collected priorto assaying the cell pellet of RIP/GLP-1 MIN6 cells byFACS analysis for Annexin-V revealed the presence of4·6�1·1 nM of GLP-1. No GLP-1 was detected inparental MIN6 cells. The concentration of GLP-1appeared within the range that in humans with type 2diabetes has been shown capable of lowering plasmaglucose levels (Todd et al. 1997)

GLP-1 inhibited the DNA fragmentation induced byISD

Activation of a calcium-dependent endonuclease is oneof the earliest detectable events denoting an irreversiblecommitment to cell death. This results in thecharacteristic formation of DNA fragments of 180–200 bp. Electrophoresis of genomic DNA from MIN6cells exposed to ISD showed the characteristic ladderingpattern that leads to cell death (Fig. 3). This was in clear

Figure 2 Evaluation of cell survival using the LIVE/DEAD viability/cytotoxicity assay:GLP-1 reduced the percentage of the cell dead induced by exposure to ISD. Parentaland RIP/GLP-1 MIN6 cells were cultured for 48 h in the presence, or absence, of ISD(sirolimus (25 ng/ml), mycophenolate (17·5 ug/ml) and tacrolimus (75 ng/ml)). Thecolor green in the cell staining identifies the living cells, while the red color identifiesthe dead cells. (A) parental MIN6 cells cultured in control medium, (B) RIP/GLP-1MIN6 cells cultured in control medium, (C) parental MIN6 cells cultured in thepresence of ISD, (D) RIP/GLP-1 MIN6 cells cultured in the presence of ISD. Thepictures shown are representative of three independent experiments which had verysimilar results. Scale bar=5µm.




contrast with the preserved DNA integrity that wasobserved in RIP/GLP-1 MIN6 cells exposed to ISD.

GLP-1 enhanced the expression of anti-apoptoticproteins and induced a down regulation ofpro-apoptotic factors

GLP-1 increased the intracellular levels of Bcl-2

RIP/GLP-1 MIN6 cells exhibited a significantly greaterabundance of the anti-apoptotic protein Bcl-2 compared

with parental cells (Fig. 4), and its abundance wasminimally affected by the exposure of cells to ISD.

GLP-1 inhibits the cleavage of the p-85 subunit of PARP

Western blot analysis revealed that in MIN6 exposed toISD there was a greater abundance of the p85 subunit ofPARP, compared with untreated cells (Fig. 4). Incomparison, RIP/GLP-1 MIN6 cells cultured in thepresence of the ISD showed a significant preservation ofthe PARP in its inactive form, and exposure of

Table 1 Quantitative analysis of the effect of GLP-1 on cell viability

Living (%) Apoptosis (%) Necrosis (%)

Cells and TreatmentMIN6 83·46±4 10·74±3 5·29±5MIN6±ISD 58·51±5 18·83±5 17·60±5RIP/GLP-1 MIN6 86·45±4 * 5·76±4 * 5·37±3 *RIP/GLP-1 MIN6+ISD 75·64±5 8·33±5 6·73±5

Percentage of living, necrotic and apoptotic cells (mean±S.E.), as derived from at least four independent experiments.Statistical significance of the data was evaluated by Student’s t test, and further confirmed by Dunnett’s post hoc test.*P,0·01.

Figure 3 GLP-1 inhibited DNA fragmentation induced by ISD. Parental and RIP/GLP-1MIN6 cells were cultured for 48 h in the presence or absence of ISD (sirolimus(25 ng/ml), mycophenolate (17·5 ug/ml) and tacrolimus (75 ng/ml)). (A) parental MIN6cells cultured in regular medium, (B) parental MIN6 cells cultured in the absence ofFBS, (C) parental MIN6 cells cultured in the absence of FBS and in the presence ofISD, (D) RIP/GLP-1 MIN6 cells cultured in the absence of FBS, (E) RIP/GLP-1 MIN6cells cultured in the absence of FBS and in the presence of ISD. Each experimentwas repeated at least three times, and the blot presented is representative of threeindependent experiments.




RIP/GLP-1 MIN6 cells did not alter the intracellularlevel of p85.

GLP-1 prevents the mitochondrial release of Smac/DIABLO

Parental MIN6 exposed to the cocktail of ISD for 48 hexhibited a significant increase in Smac/DIABLO levelscompared with untreated cultures (Fig. 4). This effectwas significantly reduced in RIP/GLP-1 MIN6 cells,even after exposure to ISD.

GLP-1 inhibits the activation of caspase-3

The exposure of parental MIN6 cells to ISD produced adetectable activation of caspase-3, as determined by

immunostaining (Fig. 5, top panel). Conversely, RIP/GLP-1 MIN6 cells were resistant to the activation ofcaspase-3 after exposure to ISD.

Western blot analysis for the active form of caspase-3further confirmed the observation derived from theimmunofluorescence study (Fig. 5, bottom panel), anddemonstrated that while in parental MIN6 cells theexposure to ISD produced a 4·5-fold increase in theintracellular level of the active form of caspase-3(ISD-treated cells vs. cells cultured with vehicle alone;P<0·001), the effect of ISD on RIP/GLP-1 MIN6 cellswas less pronounced. Indeed, RIP/GLP-1 MIN 6 cellsshowed an induction in the activation of caspase-3 of1·9 fold when compared with cells cultured withvehicle alone (P<0·01), and this was equivalent to

Figure 4 GLP-1 enhanced the expression of Bcl-2 and inhibited the expression of the pro-apoptotic proteins PARP-p85 andSmac. Parental and RIP/GLP-1 MIN6 cells were cultured for 48 h in the presence or absence of ISD (sirolimus (25 ng/ml),mycophenolate (17·5 ug/ml) and tacrolimus (75 ng/ml)). After protein extraction, the samples were subjected to western blotanalysis for Bcl-2, PARP, Smac and �-actin. (A) parental MIN6 cells cultured in control medium, (B) parental MIN6 cells cultured inthe presence of ISD, (C) RIP/GLP-1 MIN6 cells cultured in control medium, (D) RIP/GLP-1 MIN6 cells cultured in the presence ofISD. The blots on the top are representative of four independent experiments. The graphs on the bottom represent the average offour independent western blotting experiments. Protein levels were normalized for �-actin level and are shown in the graphs asfold difference, compared with control (A). *P,0·01.




approximately 60% lower than parental MIN6 cellscultured under the same conditions (ISD-treated MIN6cells vs. ISD-treated RIP/GLP-1 MIN6 cells, P<0·001).

The anti-apoptotic effect of GLP-1 was independentfrom its effect on insulin secretion

In order to investigate whether the capability ofRIP/GLP-1 MIN6 cells to counteract the pro-apoptoticaction of ISD was a direct effect of GLP-1, or whether it

was the result of a greater secretion of insulin in cellsexpressing GLP-1, we cultured cells in the presence ofdiazoxide (250 µM/L for 48 h), an agent known toinhibit insulin secretion. The viability of cells wasinvestigated using the DNA laddering assay and theircapability to secrete insulin was measured using a RIA.We demonstrated that, while diazoxide was effective insuppressing the secretion of insulin (2·36�0·21 ng ofinsulin/mg of protein in untreated cultures vs.0·41�0·07 ng of insulin/mg of protein; P<0·01), its

Figure 5 GLP-1 reduced cellular expression of active form of caspase-3. Top panels, immunofluorescence for nuclei (blue), insulin(red) and caspase-3 (green). A–C, immunostaining of parental MIN6 cells; A, Hoechst nuclear staining; B, insulin; C, caspase-3;D—F, immunostaining of RIP/GLP-1 MIN6 cells; D, Hoechst nuclear staining; E, insulin; F, caspase-3; G–I, immunostaining ofparental MIN6 cells (cultured in the presence of ISD; sirolimus (25 ng/ml), mycophenolate (17·5 ug/ml) and tacrolimus (75 ng/ml)for 48 hours); G, Hoechst nuclear staining; H, insulin; I, caspase-3; J–L, immunostaining of RIP/GLP-1 MIN6 cells cultured in thepresence of ISD; J, Hoechst nuclear staining; K, insulin; L, caspase-3. Scale bar 5µ. Each experiment was repeated at least threetimes, and the pictures shown are representative of three independent experiments. Bottom panels, western blot analysis for theactive form of caspase-3. Parental and RIP/GLP-1 MIN6 cells were cultured for 48 h in the presence or absence of ISD. Afterprotein extraction, the samples were subjected to western blot analysis for the active form of caspase-3 and �-actin. A, parentalMIN6 cells cultured in control medium; B, parental MIN6 cells cultured in the presence of ISD; C, RIP/GLP-1 MIN6 cells cultured incontrol medium; D, RIP/GLP-1 MIN6 cells cultured in the presence of ISD. The blot on the left is representative of threeindependent experiments. The graph on the right represents the average of three independent experiments. Protein levels werenormalized for �-actin level and are shown in the graphs as fold difference, compared with control (A). *P,0·01; **P,0·001.




action did not diminish the resistance of RIP/GLP-1MIN6 cells to the pro-apoptotic action of ISD (Fig. 6,lanes D vs. E).

Parental MIN6 cells exposed to ISD showed thecharacteristic laddering pattern that leads to cell death(Fig. 6, lane C). In contrast, RIP/GLP-1 MIN6 cells didnot show DNA fragmentation when exposed to ISD.These results supported the hypothesis that theantiapoptotic role of GLP-1 on the cells exposed to acocktail of ISD was independent from its effect on thestimulation of insulin secretion.

Inhibition of GLP-1 receptor activation by exendin 9abolished the effect of GLP-1 on the viability of cellsexposed to ISD

To investigate whether the capability of RIP/GLP-1MIN6 cells to counteract the proapoptotic effect of ISDwas specifically mediated by the activation of the GLP-1receptor, cells were cultured in medium containingincreasing concentrations of the receptor antagonistexendin-9. FACS analysis for Annexin V demonstratedthat 19·60�2% of parental MIN6 cells vs. 8·34�2·3%of RIP/GLP-1 MIN6 cells were Annexin-V-positivewhen cultured for 48 h in medium containing ISD(P<0·01) (Fig. 7). The cell survival advantage thatRIP/GLP-1 MIN6 cells had when compared withparental MIN6 cells was abolished by the presence ofexendin 9 in the culture medium. Indeed, as theconcentration of exendin 9 in the medium wasincreased, the difference in the percentage of living cellsbetween the two cell lines (RIP/GLP-1 MIN6 cells andparental MIN6 cells) became no longer significant. Thecapability of exendin-9 to inhibit cell death reached itsplateau at the concentration of 10 nM, and at thatconcentration the percentage of death cells was verysimilar among cultures (43·4�2·1 vs. 41·7�1·9,in parental vs. RIP/GLP-1 MIN6 cells, respectively)(Fig. 7).

These results demonstrate that the inhibition ofthe activation of the GLP-1 receptor abolish thecapability of RIP/GLP-1 MIN6 cell to counteract thepro-apoptotic action of ISD.

GLP-1 induced a dose-dependent inhibition of celldeath

To investigate the dose-dependent inhibition of celldeath by GLP-1 we cultured parental MIN6 cells in thepresence of various concentration of GLP-1 (for 48 h)and then exposed each culture condition to ISD (for anadditional 48 h). In addition to defining the doseresponse for GLP-1�s effect on cell viability, the aim ofthis experiment was to investigate whether the observedcapability of RIP/GLP-1 MIN6 to be protected againstthe toxic effect of ISD was due to the specific action of

GLP-1, or whether it was the result of a clonal selectionthat rendered the transfected cell line less susceptible tocell death. FACS analysis for Annexin-V demonstratedthe presence of a dose-dependent inhibition of cell deathby GLP-1 (Fig. 8). The presence of 1 nM GLP-1 in theculture medium was sufficient to produce a statisticallysignificant inhibition of cell death (38% vs. 29·7% incontrol vs. GLP-1 treated cultures, respectively; P<0·01)and it further increased in a dose-dependent fashion incultures grown in 5, 10 or 100 nM of GLP-1.

Discussion

In the present study we demonstrated that some of themost important immunosuppressive drugs used afterislets transplantation have a negative effect on functionand viability of insulin-producing cells. The presence ofimmunosuppressive agents in the culture medium ofMIN6 cells altered their glucose-dependent capability tosecrete insulin, and caused a significant increase in thenumber of cells undergoing apoptosis and necrosis. Thiscomplex set of biological effects was significantly reducedby the transfection of cells with a plasmid expressing theGLP-1 fragment of the human proglucagon gene.

Our findings are consistent with previous reportsdescribing deterioration in islet function and viability inresponse to tacrolimus, sirolimus and mycophenolate.Drachenberg et al. (1999) performed electron microscopystudies of pancreatic biopsies from patients subjected topancreas and kidney transplantation and showed thattreatment with tacrolimus was associated with anincreased cytoplasmic swelling, cellular vacuolization,and a marked decrease in dense-core secretory granulesof islet cells. Paty et al. (2002) demonstrated thatmycophenolate and tacrolimus had deleterious effects oninsulin secretion in both HITT15 cells and cultured ratislets. In addition to inhibiting insulin secretion (Ricordiet al. 1991, Strasser et al. 1992, Ishizuka et al. 1993),tacrolimus has been shown to down regulate thesynthesis of insulin (Redmon et al. 1996). Mycophenolatehas been shown to reduce insulin secretion by inhibitingvoltage-dependent calcium channels (Meredith et al.1997, Li et al. 2000). On the other hand, sirolimus hasbeen shown to have a very modest adverse effect on isletfunction (Fabian et al. 1992, Kneteman et al. 1995); and areduction in insulin release has been demonstrated onlyafter a prolonged culture with a high concentration ofthe drug (Fabian et al. 1992). Prior observations fromvarious laboratories have shown that GLP-1 was capableof inducing the expression of anti-apoptotic proteinsleading to improved survival of cells. Li et al. (2003) haveshown that GLP-1 receptor signaling modifies thesusceptibility of mice to undergo islet cell apoptosis aftertreatment with streptozotocin. Farilla et al. (2002) havedemonstrated that GLP-1 reduced the number of




Figure 6 The anti apoptotic effect of GLP-1 is not blocked by the inhibition of insulin secretion.Parental and MIN6 RIP/GLP-1 cells were cultured for 48 h in the presence or absence of diazoxide(250µM/L) and in the presence or absence of ISD (sirolimus (25 ng/ml), mycophenolate (17·5 ug/ml)and tacrolimus (75 ng/ml)). On the top, DNA fragmentation assay. A, parental MIN6 cells cultured inregular medium; B, parental MIN6 cells cultured in the presence of diazoxide; C, parental MIN6cultured in the presence of diazoxide and in the presence of ISD; D, RIP/GLP-1 MIN6 cells cultured inthe presence of diazoxide; E, RIP/GLP-1 MIN6 cells cultured in the presence of diazoxide and in thepresence of ISD. On the bottom, insulin secretion assay. A, parental MIN6 cells cultured in regularmedium; B, parental MIN6 cells cultured in the presence of diazoxide; C, parental MIN6 cultured inthe presence of diazoxide and in the presence of ISD; D, RIP/GLP-1 MIN6 cells cultured in thepresence of diazoxide; E, RIP/GLP-1 MIN6 cells cultured in the presence of diazoxide and in thepresence of ISD. Each experiment was repeated at least three times and the data plotted on thegraph represent the mean±S.E. *P,0·01




Figure 7 Inhibition of GLP-1 receptor activation by exendin 9 abolished the beneficial effect of GLP-1on the viability of cells exposed to ISD. Parental and MIN6 RIP/GLP-1 cells were cultured for 48 h inthe presence of ISD (sirolimus (25 ng/ml), mycophenolate (17·5 ug/ml) and tacrolimus (75 ng/ml)) andexposed to different doses of exendin 9. The graph depicts the percentage of apoptotic and necroticparental and RIP/GLP-1 MIN6 cells. Each experiment was repeated at least four times and the dataplotted on the graph represent the mean±S.E. *P,0·01, # P,0·05.

Figure 8 Dose-dependent inhibition of cell death by GLP-1. Parental MIN6 cells were cultured for48 h in the presence of various concentration of GLP-1 (0, 0·1, 0·5, 1, 5, 10, 100 mM). A 10 µl aliquotof GLP-1 was added every 8 h, without changing the culture medium. After 48 h, the old medium wasdiscarded and replaced with fresh medium containing various concentration of GLP-1 in the presenceof ISD (sirolimus (25 ng/ml), mycophenolate (17·5 ug/ml) and tacrolimus (75 ng/ml)) for an additionalperiod of 48 h. At the end of this second incubation period the medium was discarded and the pelletassayed for cell viability by FACS analysis for Annexin-V. The graph depicts the percentage ofapoptotic and necrotic cells. Each experiment was repeated at three times, and the data plotted onthe graph represent the mean±S.E *P,0·01, ** P,0·001.




apoptotic cells in the pancreas of Zucker diabetic rats,and that GLP-1 when added to freshly isolated humanislets or cultured MIN6 cells preserved cell morphology,enhanced insulin synthesis, and improved the cellularcapability of secreting insulin in response to glucose(Farilla et al. 2003, Hui et al. 2003). Interestingly, aprotective effect of GLP-1 from cellular apoptosis hasalso been shown on cells other than the pancreaticendocrine cells, as for the neuronal cells of hippocampus(During et al. 2003).

In the present report, studying factors that maymediate the anti-apoptotic action of GLP1, weinvestigated its effect on some of the main regulators of

cell apoptosis, including Bcl-2, caspase-3, PARP, andSmac.

Bcl-2 is member of a class of proteins that are mainlylocalized at the outer mitochondrial membrane and itsexpression is associated with inhibition of cell apoptosis.Our study demonstrates that in �-cells transfected with aplasmid encoding for GLP-1 there was a constitutiveactivation of the expression of Bcl-2. This observation isconsistent with studies by Contreras (2001) andRabinovitch (1999) showing that the overexpression ofBcl-2 confers a long-term protection of islet mass aftertransplantation (Rabinovitch et al. 1999, Contreras et al.2001).

Figure 9 Schematic of ISD effect in parental and GLP-1 transfected MIN6 cells. The exposure of parental MIN6 cells toimmunosuppressive agents is associated with the release of Smac/Diable from the mitochondria, an increased detection of theactive form of caspase-3 and increased production of the p85 subunit of PARP (left panel). The transfection of MIN6 cells with theGLP-1 fragment of the proglucagon gene is associated with an increased expression of Bcl-2, a decreased cytosolic level ofSmac, a down regulation of the levels of the active form of caspase-3 and a decreased abundance of the p85 fraction of PARP(right panel).




All apoptotic pathways, so far identified, convergetoward the activation of cytoplasmic cystine proteasesnamed caspases and of the various caspases identified,caspase-3 is one of those that better correlates with cellapoptosis. PARP is involved in the regulation of variousimportant cellular processes such as differentiation,proliferation, and tumor transformation and also in theregulation of the molecular events involved in therecovery of cells from DNA damage (Uchida et al. 2001).In our study, we demonstrated that the exposure ofMIN6 cells to immunosuppressive agents was associatedwith an increased detection of the active form ofcaspase-3 and an increased production of the p85subunit of PARP. In cells transfected with GLP-1 theexpression of the active form of caspase-3 and ofPARPp85 were significantly reduced (Fig. 9).

GLP-1 also inhibited the release of Smac/DIABLO, adimeric mitochondrial protein released into the cytosolalong with cytochrome-c during the execution phase ofthe intrinsic pathway of apoptosis. Because the release ofSmac has been proposed to directly enable caspase-3activation, its detection is an additional indication that adeath signal is activated and its down regulation in cellstransfected with GLP-1 is a further evidence of anincrease resistance against the pro-apoptotic action ofimmunosuppressive drugs.

Finally, we demonstrated that the capability of GLP-1transfected MIN6 cells to be resistant to the pro-apoptotic action of immunosuppressive agents wasdirectly mediated by the autocrine production of GLP-1and it was not due to its effect on the secretion of insulin.

In summary, our results suggest that GLP-1 is apowerful antiapoptotic agent capable of protecting cellsfrom damage induced by immunosuppressive drugs.These observations may have important clinical andtherapeutic implications as GLP-1 is being considered asa potential pharmacological agent for the treatment ofdiabetes.

Acknowledgments

This study was supported, in part, by the Max FactorFamily Foundation.

References

Bennet W, Sundberg B, Groth CG, Brendel MD, Brandhorst D,Brandhorst H, Bretzel RG, Elgue G, Larsson R, Nilsson B &Korsgren O 1999 Incompatibility between human blood andisolated islets of Langerhans: a finding with implications forclinical intraportal islet transplantation? Diabetes 48 1907–1912.

Contreras JL, Bilbao G, Smyth CA, Jiang XL, Eckhoff DE, JenkinsSM, Thomas FT, Curiel DT & Thomas JM 2001 Cytoprotectionof pancreatic islets before and soon after transplantation by genetransfer of the anti-apoptotic Bcl-2 gene. Transplantation 711015–1023.

Drachenberg CB, Klassen DK, Weir MR, Wiland A, Fink JC,Bartlett ST, Cangro CB, Blahut S & Papadimitriou JC 1999 Isletcell damage associated with tacrolimus and cyclosporine:morphological features in pancreas allograft biopsies and clinicalcorrelation. Transplantation 68 396–402.

Drucker DJ 2003 Glucagon-like peptides regulators of cellproliferation, differentiation, and apoptosis. Molecular Endocrinology17 161–171.

During MJ, Cao L, Zuzga DS, Francis JS, Fitzsimons HL, Jiao X,Bland RJ, Klugmann M, Banks WA, Drucker DJ & Haile CN2003 Glucagon-like peptide-1 receptor is involved in learning andneuroprotection. Nature Medicine 9 1173–1179.

Fabian MC, Lakey JR, Rajotte RV & Kneteman NM 1992Rapamycin prolongs murine islet al.lograft survival. TransplantationProceedings 24 2842.

Farilla L, Hui H, Bertolotto C, Bulotta A, Kang E, Di Mario U &Perfetti R 2002 GLP-1 promotes islet cells growth and inhibits cellapoptosis in Zucker diabetic rats. Endocrinology 143 4397–4408.

Farilla L, Bulotta A, Hirshberg B, Li Calzi S, Khoury N,Noushmehr H, Bertolotto C, Di Mario U, Harlan DM & PerfettiR 2003 Glucagon-like peptide 1 inhibits cell apoptosis andimproves glucose responsiveness of freshly isolated human islets.Endocrinology 144 5149–5158.

Hering B & Ricordi C 1999 Islet transplantation for patients withtype 1 diabetes. Graft 2 12–27.

Herold KC, Nagamatsu S, Buse JB, Kulsakdinun P & Steiner DF1993 Inhibition of glucose-stimulated insulin release from betaTC3 cells and rodent islets by an analog of FK506. Transplantation55 186–192.

Hui H, Yu R, Bousquet C & Perfetti R 2002 Transfection ofpancreatic-derived �-cells with a minigene encoding for humanglucagon-like peptide-1 regulates glucose-dependent insulinsynthesis and secretion. Endocrinology 143 3529–3539.

Hui H, Nourparvar A, Zhao X & Perfetti R 2003 Glucagon-LikePeptide-1 Inhibits Apoptosis of Insulin-Secreting Cells via a Cyclic5’-Adenosine Monophosphate-Dependent Protein Kinase A-and aPhosphatidylinositol 3-Kinase-Dependent Pathway Endocrinology144 1444–1455.

Hui H, Khoury N, Zhao X, D’ Amico E, Bulotta A, Nguyen E &Perfetti R 2004 Adenovirus-mediated XIAP gene transfer reversesthe negative effects of immunosuppressive drugs on insulinsecretion and viability of isolated human islets and MIN6 cells.64th Annual Meeting of American Diabetes Association, Orlando,FL.

Hussain MA & Habener JF 2000 Glucagon-like peptide 1 increasesglucose-dependent activity of the homeoprotein IDX-1transactivating domain in pancreatic �-cells. Biochemical andBiophysical Research Communications 11 616–619.

Ishizuka J, Gugliuzza KK, Wassmuth Z, Hsieh J, Sato K,Tsuchiya T, Townsend CM Jr, Fish JC & Thompson JC1993 Effects of FK506 and cyclosporine on dynamic insulinsecretion from isolated dog pancreatic islets. Transplantation 561486–1490.

Kneteman NM, Lakey JR, Wagner T & Finegood D 1995 Beneficialmetabolic impact of the novel immunosuppressant rapamycin inchronic canine islet autograft recipients. Transplantation Proceedings27 3213–3217.

Lakey JR, Warnock GL, Shapiro AM, Korbutt GS, Ao Z,Kneteman NM & Rajotte RV 1999 Intraductal collagenasedelivery into the human pancreas using syringe loading orcontrolled perfusion. Cell Transplant 8 285–292.

Li GD, Luo RH & Metz SA 2000 Effects of inhibitors of guaninenucleotide synthesis on membrane potential and cytosolic freeCa2+ levels in insulin-secreting cells. Biochemical Pharmacology 59545–556.

Li Y, Hansotia T, Yusta B, Ris F, Halban PA & Drucker DJ 2003Glucagon-like peptide-1 receptor signaling modulates � cellapoptosis. Journal of Biological Chemistry 278 471–478.




Meredith M, Li G & Metz SA 1997 Inhibition of calcium-inducedinsulin secretion from intact HIT-T15 or INS-1 beta cells by GTPdepletion. Biochemical Pharmacology 53 1873–1882.

Patty BW, Harmon JS, Marsh LC & Robertson RP 2002Inhibitory effect of immunosuppressive drugs on insulinsecretion from HIT-T15 cells and Wistar rat islets. Transplantation353–357.

Rabinovitch A, Suarez-Pinzon W, Strynadka K, Ju Q, Edelstein D,Brownlee M, Korbutt GS & Rajotte RV 1999 Transfection ofhuman pancreatic islets with an anti-apoptotic gene (Bcl-2)protects beta-cells from cytokine-induced destruction. Diabetes 481223–1229.

Redmon JB, Olson LK, Armstrong MB, Greene MJ & RobertsonRP 1996 Effects of tacrolimus (FK506) on human insulin geneexpression, insulin mRNA levels, and insulin secretion inHIT-T15 cells. Journal of Clinical Investigation 98 2786–2794.

Ricordi C, Tzakis A, Alejandro R, Zeng YJ, Demetris AJ, Carroll P,Mintz DH & Starzl TE 1991 In vivo effect of FK506 on humanpancreatic islets. Transplantation 52 519–522.

Sato T, Inagaki A, Uchida K, Ueki T, Goto N, Matsuoka S,Katayama A, Haba T, Tominaga Y, Okajima Y, Ohta K, SugaH, Taguchi S, Kakiya S, Itatsu T, Kobayashi T & Nakao A 2003Diabetes mellitus after transplant: relationship to pretransplantglucose metabolism and tacrolimus or cyclosporine A-basedtherapy. Transplantation 76 1320–1326.

Shapiro AM, Lakey JR, Ryan EA, Korbutt GS, Toth E, WarnockGL, Kneteman NM & Rajotte RV N 2000 Islet transplantation inseven patients with type 1 diabetes mellitus using a

glucocorticoid-free immunosuppressive regimen. New EnglandJournal of Medicine 343 230–238.

Strasser S, Alejandro R, Shapiro ET, Ricordi C, Todo S & MintzDH 1992 Effect of FK506 on insulin secretion in normal dogs.Metabolism 41 64–67.

Sutherland DE, Gruessner RW, Dunn DL, Matas AJ, Humar A,Kandaswamy R, Mauer SM, Kennedy WR, Goetz FC, RobertsonRP, Gruessner AC & Najarian JS 2001 Lessons learned frommore than 1000 pancreas transplants at a single institution. Annalsof Surgery 233 463–501.

The Diabetes Control and Complications Trial Research Group1998 Effect of intensive therapy on residual �-cell function inpatients with type 1 diabetes in the diabetes control andcomplications trial: a randomized, controlled trial. Annals of InternalMedicine 128 517–523.

Todd JF, Wilding JP, Edwards CM, Khan FA, Ghatei MA & BloomSR 1997 Glucagon-like peptide-1 (GLP-1): a trial of treatment innon-insulin-dependent diabetes mellitus. European Journal of ClinicalInvestigation 27 533–536.

Uchida M, Hanai S, Uematsu N, Sawamoto K, Okano H, Miwa M& Uchida K 2001 Genetic and functional analysis of PARP, aDNA strand break-binding enzyme. Mutation Research 477 89–96.

Zeng Y, Ricordi C, Lendoire J, Carroll PB, Alejandro R, BereiterDR, Tzakis A & Starzl TE 1993 The effect of prednisone onpancreatic islet autografts in dogs. Surgery 113 98–102.

Received 26 November 2004Accepted 8 December 2004




Pancreatic -cells expressing GLP-1 are resistant to the toxic effects of immunosuppressive drugs

Documents